A lawsuit filed July 21 claims that ChatGPT offered inaccurate medical guidance to a Florida man that resulted in delayed care for a pulmonary embolism last year.
The lawsuit, filed in the Superior Court of California in San Francisco, accuses OpenAI and CEO Sam Altman of negligence, unlicensed practice of medicine, defective product design and other claims. It is among the first to allege that a chatbot’s health advice caused physical injury to a user seeking guidance about a medical condition, according to The New York Times, and the plaintiff has asked the court to stop ChatGPT Health from operating until independent evaluators verify its safety.
According to the complaint, ChatGPT-4o told Scott Winters, a pastor, that early signs of his health problems were “not something dangerous” and discouraged him from seeking medical care, urging him instead to trust that “God did not design your body to endlessly fail.” Mr. Winters, 55, claims ChatGPT manipulated his own language and beliefs with the knowledge that he was a pastor.
He began using ChatGPT in 2024 for everyday questions and general health queries. Over time, the lawsuit claims, the chatbot’s disclaimers urging him to seek professional care fell away, and it began offering direct guidance instead — including, when he described dizziness and blood pressure instability in June 2025, advice to stay home and “recliner-bound.” Mr. Winters followed the guidance and limited his movement.
“ChatGPT-4o positioned itself as an authoritative partner in a medical journey it was simultaneously designing for Scott,” alleges the complaint, which includes screenshots of Mr. Winters’ exchanges with the chatbot.
In July 2025, Mr. Winters turned to the chatbot when he developed pain in his groin, which ChatGPT dismissed as “very likely another minor piece of the long story” involving previously discussed health issues with his gut and pelvic floor. The next morning, Mr. Winters was hospitalized in the ICU with a massive pulmonary embolism due to blood clots in both lungs, which physicians attributed to weeks of immobility, according to the complaint.
Drew Pusateri, an OpenAI spokesperson, told The New York Times that ChatGPT’s terms of service state that the chatbot is not meant to be used for medical diagnosis or treatment. Even so, he said the company takes seriously its responsibility to make its answers to health questions as safe as possible, since people do turn to the technology for that purpose.
“Treating chatbots as the whole story behind people’s medical decisions or outcomes oversimplifies a much bigger challenge, and risks getting in the way of people accessing powerful new tools that can aid them in their health journey,” Mr. Pusateri said in a statement to The Times. He added that OpenAI’s newer models are an improvement on the one Mr. Winters used, with newer models that seek missing information more effectively, signal uncertainty and recognize when someone should seek professional care.
The lawsuit seeks monetary and punitive damages, a pause on ChatGPT Health pending independent safety audits, and injunctive relief requiring OpenAI to hard-code refusals for medical diagnosis and treatment questions and to automatically end conversations when emergency care is needed. Mr. Winters is represented by Tech Justice Law and the Social Media Victims Law Center.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.